Experimental Design Example

An operator running a process usually cannot just by looking tell what the optimum setting for a process is without some sort of statistical analysis. Here is a record of the machine settings and the output that came out for those settings.

By looking at these numbers, what is the best setting to minimize the Output numbers?

Let’s say you are the operator. By eyeballing the data here, what should you do to minimize Output?

Many people who know a little bit of statistics would choose a correlation analysis to answer this question. Click here to see what a correlation tells you.

Some people might tell you to do a linear regression. Click here to see what a linear regression would tell you.

 

Answer

 
Setting A Setting B Output

13.8
17.5
7.0
18.5
11.3
16.4
12.8
22.9
8.3
22.5
9.4
16.6
19.4
13.6
15.2
19.6
10.7
20.2
14.4
14.8
9.3
17.5
21.8
24.2
11.2
23.5
20.0
17.3
20.2
8.9
20.8
7.2
24.9
14.3
16.8
21.9
9.2
9.2
12.5
14.8

7.7
13.9
21.0
13.6
15.3
19.8
14.3
13.4
7.9
14.5
14.2
6.2
23.7
22.9
24.3
19.5
6.1
19.3
8.1
22.5
9.6
17.7
14.4
8.8
22.8
17.9
15.1
14.8
11.8
9.1
21.2
12.7
19.7
24.2
18.7
10.4
16.2
24.9
13.7
13.7

14.4
15.6
18.1
14.2
16.2
14.9
15.2
16.8
16.8
15.1
13.5
14.0
14.3
16.0
16.0
14.1
12.5
13.9
16.3
14.1
16.5
10.2
14.0
18.5
14.4
13.2
16.9
15.4
16.5
13.4
14.0
16.4
11.9
15.8
15.9
14.5
12.0
18.8
12.6
15.4